Simon Watson: The Thinking Behind SongSyntax | Creator Spotlight
In October, Simon Watson took AI out of the step that builds SongSyntax prompts. That decision offers a window into how he builds tools for musicians, and how the work led him into games.
Before the first note is generated, somebody has to make the musical decisions. What interests me about Simon Watson’s work is how much of that decision-making happens before you hit Generate.
I first featured Simon in July 2026, when SongSyntax was still taking shape. I wanted to see whether a more structured brief could help me make clearer decisions about a song. That question matters even more now that Simon has changed how the tool builds those briefs.
Why Simon built a different kind of prompt workflow
A person may know exactly how a song should feel and still struggle to translate that feeling into musical instructions. “Make it more dramatic” or “give it a Baroque quality” can describe a real creative intention without settling the harder choices: which instruments carry the idea, how sections should unfold, what period conventions matter, or which conflicting sounds should be excluded.
Simon approached that gap through SongSyntax, researching musical styles and building prompts around decisions a creator can actually make. If you’re asking for an instrumental, for instance, you can decide which instrument carries the melody and when the arrangement should open up.
The October decision that changed SongSyntax
In early October, Simon says he moved SongSyntax from AI-based prompt generation to an engine that assembles prompts without a generative AI model. SongSyntax builds the musical instructions. A separate AI music generator still turns those instructions into audio.
Simon says the change was driven by both consistency and cost. In his earlier tests, using an AI model to generate a prompt could add a per-generation expense that varied with complexity. His current approach avoids that particular AI-generation step in SongSyntax. It does not mean the music generator or every SongSyntax product is free.
A repeatable prompt structure gives creators a way to work with the same musical choices from one attempt to the next.
What I find interesting is where Simon chose to leave room for variation. If I change the instrumentation in a brief, I want to hear what that decision does to the track. A prompt with a consistent structure gives me a way to keep track of what I asked for, while the music generator can still surprise me with what it does with those instructions.
Where Spectrum and Suno hand the work back to each other
Simon gave me another useful detail after reviewing this draft on November 8. In SongSyntax Spectrum, a creator can write their own lyrics into the structure the tool provides. Or they can use that structure as a starting point for lyrics generated in Suno.
The workflow Simon described · November 2026
- Start in Spectrum. Build a Lyrics/Structure section around the musical idea, using your own lyrics if you already have them.
- Use Suno Simple mode if you want lyric suggestions. Paste that Lyrics/Structure section into Suno's Simple mode and let Suno generate lyrics.
- Return to Spectrum. Bring those lyrics back so SongSyntax can insert them into its structured output.
- Finish in Suno's Advanced mode. Take the prepared sections back to Suno and work separately with Lyrics/Structure, Style and Exclusions.
This is Simon's description of the workflow as it stood on November 8, rather than a claim that I have independently retested the current Spectrum-to-Suno handoff. Tool interfaces can change; for the latest instructions, the creator's own site is the right starting point.
That distinction matters. SongSyntax is arranging the instructions with code and curated material. When a creator chooses to generate lyrics or music in Suno, that is where the generative AI comes in. The process still leaves the creator responsible for deciding what to keep, rewrite or exclude.
What I learned from an early SongSyntax test
Before the October change, I used SongSyntax alongside Suno to see what a guided brief could do for a track. On my first try, I’d already made nearly every production decision. The more informative experiment came when I started with the purpose and emotional direction of the song, then left the structure and musical relationships room to develop.
That process fed a series of FIRE PON ROME experiments, including an unexpected Baroque direction. The version I liked best came from listening back, keeping what worked and making the final choices myself.
More than SongSyntax: writing about AI responsibility
Simon’s books explore a question I also see in SongSyntax. Where can automation help, and which decisions still belong to the person using it?
In The Sorcerer's Apprentice at Work and its Essential Guide, Simon examines questions of AI risk, governance, evidence and oversight. In LYTIC, he turns to questions about change and dissolution. These are different subjects from AI music, but they connect through a familiar decision: what can technology do usefully, and where must people remain accountable?
I've reviewed these works separately because each deserves an honest examination on its own terms. They also give this Creator Spotlight a much better sense of the breadth of Simon's interests than a single product announcement could.
Simon’s own music
In Falstaff Is Dead, Simon takes a question from Shakespeare into his music. What happens to an old friendship when one person comes into power? That’s the part that interests me as a songwriter. A question like that gives you something to listen for when you’re deciding which musical ideas to keep.
Simon also releases music as We Ain’t Real on Suno. His Suno profile gives readers a place to hear his work alongside the tools and books he’s built.
How SongSyntax led Simon into games
Simon traces his Field Commander games back to what he learned building SongSyntax. The projects range from Napoleonic strategy to cricket, baseball and football. He also sees a way for the two projects to work together. SongSyntax can help shape a musical brief for a game soundtrack, giving the tools he built for music creators a place in his own game work.
What ties the projects together for me is the path he took from one to the other. Building a system around musical decisions gave him a way into a different kind of creative work.
Simon makes a useful distinction between the two lines of work. He says many Field Commander games are free or inexpensive, with tips welcome, because he wants people to play them. SongSyntax is a separate family of music-creator tools, built through extensive work and intended to have its own commercial value. I would not treat one as a free version of the other.
What I’d take into the next session
For someone just beginning with AI music, the practical lesson is simple: know what you're trying to make before asking a system to make it. Use tools to explore structure and possibilities. Then listen, compare, reject and revise. Keep the judgment where it belongs.
